Nonsub - Robotics Question

Annual Capital Expenditures Survey

OMB Clearance Number 0607-0782

Approval Date 12/21/2018 covering FY19- FY21

Nonsubstantive Change for FY19 - Form ACE-1

Beginning with FY2019, the Census Bureau plans to collect expenditures for both industrial and service robotics at the company level across 19 NAICS sectors. These expenditures will be collected from a sample of 50,000 businesses with employees on the 2018 Annual Capital Expenditures Survey (ACES). The new question is item 5B on the ACE-1(L), ACE-1(M) and ACE-1(S) forms.

We submitted the robotics question for formal OMB review in October 2018. However, the question was still undergoing cognitive testing at that time. We were instructed by OMB to resubmit the robotics question for review as a non-substantive change once testing was complete and final wording developed.

The Census Bureau conducted two rounds of debriefing interviews in person and over the phone with respondents of 26 companies from April 1, 2018 through mid-December 2018. All of the respondents interviewed had completed the 2017 ACES. The types of companies interviewed included Accommodation and Food Services, Construction, Health Care and Social Assistance, Manufacturing, Mining, Retail Trade, Transportation, and Utilities.

The goals of the cognitive testing were to determine whether respondents understood what data was being requested and whether respondents had the means to answer the question. Based on the findings and recommendations from the first round of 15 interviews conducted from April through August, modifications were made to the question, instructions and definitions. The project team members also consulted with a professor at the University of Maryland’s Robotic Center.

A second round of testing was conducted with respondents of 11 companies in November and December. The revised question, instructions, and definitions were generally understood by the respondents. The majority of respondents said that they could gather the appropriate records to answer the question and most would consult with other staff at their company for assistance. Respondents’ estimates of burden associated with acquiring the data to answer the question varied by company, from a low of 10 minutes to a high of 10-20 hours. This agrees with our initial estimate.

Recommendations from the interviews were to send advance notification of the new data request for robotic equipment expenditures on the 2018 ACES. This advance notice will be sent via email. This may help respondents determine what equipment in the company may be considered robotic and identify how these items are categorized within their records.

Respondents will be asked to provide a value or an explanation if no value can be provided. We propose to use the Remarks section for the first year or two to help evaluate responses to this question.